Europeun ,Iou;'nal o]' Pharmacok; W ~ Ah~h'cuNr Phatvnacolok, y Section, 227 (1992) 205-214

2(15

a'! 1¢,~92Elsevier Science Publishers B.V. All rights reserved 0922-411)6/92/$95.tlg

EJPMOI.. t~11357

Site-directed mutagenesis of the human dopamine

D 2

receptor

Alfred Mansour ", Fan M e n g ", James H. M e a d o r - W o o d r u f f ~', Larry P. Taylor ~', Olivier Civelli ~' and Huda A k i l '' " ,',Iomd Ih,alth Research Iretittle, UnirersiO" of ?~lichi.~an. Aml Ad~or. MI 48109-0720. USA am/ i, |.ollum [n~ttttm. ot'/tdcanced Biolo,eh al Research. Oregon Ih'ahh Sciem'c.~ Unirer~iO'. I'm'thmtL OR ~)7201, UAA

Received 15 April 1992, revised MS received 9 June 1992, accepled 23 June ItJ92

Based on "mlim) acid sequence and computer modeling, two conflicting three-dimensional mndcls of die dopamine l)~ receptor have beth proposed. One m~del (Dahl et al., 1991, Prec. Natl. Acad. Sci. USA 88, 8111) stlggcsts that dopamine interacts with asp'~"latc 80 of transmcmbranc (TM) 2 and asparaginc 390 of TM6 with the transmenlbranes arranged in a clockwise manner, white a second m,,del (Hiberl et al., 1991, Mol. Pharmacol. 40, 8) suggests that dopaminc interacts with aspartatc 114 of TM3 and the serines of TM5 (194 and 197) with the transmcmbranes arranged m a countcrclockwise manner when viewed from the cxtracellular space. The present study tests the latter model by selectively mutating aspartate 114 and serines 194 and !c-~7,af the t;uman &~pamine D, receptor by site-directed rnutagcnesis. In addition, two methioeincs (116 and 117) were mutated to evaluate whether residues near aspartate (1i4) of the dopaminc D 2 receptor are critical in differentiating dopamine receptor agonists from adrenoceptor agonists. Removal of the negative charge with the mutation of aspartatc (114) to either asparagine or glycinc led to a to!al loss ¢~fboth agonist and antagonist binding, lndNidual or dual methinnine mutations in positions 116 and 117, to make the dopamine D 2 binding pocket more closely resemble the/,12-adrenoceptor, did not result ill a change in selectivity toward noradrenergic ag~mists or anla[,,onists. Tile scrine mutations revealed interesting differences between the dopamine D 2 receptor and the adrenoceptors. In particular serine 197 appeared more imp,~rtant than serine 194 for agonist binding. In addition, the binding of Olle agonist (NdH.37) was un,J'fcctcd by individual scrinc nlutathms, white the binding of some antagonists, such as raclopride and spip,.'rone, was significantly al,.cred. These fi,ldings are discussed in relation to ligand slrucltlrc and their interactions with the putative binding pocket. 1)opamine 1)~ receptors: Mut~,gcnesis: Receptor binding: Catecholaminc receptors

1. lntro,;uctiol~,, The iecent cloning (Bunzow et al., 1988; Dearry ct al., It)90; Mahan et al., 199{); Monsma et al., 1990; Soko]off el al., 1990; Sunahara et al., 19'~fl. 1991; Z h o u el a!., !99(t; Van Tot c t a l . , 19nl) of the dopamin¢ receptors ( D ~ - D 5) suggests that they are members of a larger family of G-protein coupled, seven transmembrane receptors (Gilman, 1987) whose other members include the adrenneeptor (Dixon et al., 1986; Friellc ct al., 1987; Kobilka c t a l . , 1987a,b; C'ltecchia et al., 1988; Rcg;m et a[., 1988), muscarinic acetylcholine receptor (Bonner et al., 1987; Peralta et al., 1987: Ramacimndran ct al., 1989), and peptidergic reccoiors (Masu el al., 1987; Hershey and Krause, 19911; Tanaka ct al., 199{)). These receplors share a common structural too-

Correspoiatencc to: l)r. Alfred Mansuur. Menla'. l lcalth Research Institute. University of ,M,ichig;tp., 205 Zina Pitcher P',acc Ann Arbur, MI 4,~1(19ql720, LISA. "1"ct.(313) q3fl-21}41:Fax ~313) 747-4130.

tif and arc similar to the better characterized rhodopsir, protein (Henderson el ill., 1990)which spans the ph, snla membra~'e :;c,,en tim'?:;; widi the transmcmbrane (TM) domains forming a binding pocket. The N-termina! region is thought lo be extracellular and the C O O H terminus intracellular. Three intracelh:l.ir and threc cxtracellular loops are lormcd whh tiffs organization and the second and third cytoplasmic loops are lhougtlt to be involved iv. G-protein cnupling (Kobilka et al., 1988; Str.'.:der et al., 198"7.:). Site-directed mutagencsis studies of tiw /le-ad,enoceplor suggest that specific amino acids deep in the binding pocket formed by the TM domains arc critical for the recognition of catecholamines (Fraser, 1980: Strader et al., 1987, 1989a,b; Neve et al., 1991) In particular, aspartatc 79 in Tr',]2 and asparlat_c 113 in TM3 arc tht;'.~ght to in{era i wilh the amino group of epinephrine and norcpinc:,hrine (Strader et al., l')[(7b). wk'ile two serincs (2/14 and 2U7) in T M 5 form hydrogen bonds with the hydroxyl groups of the caicd,ol rnoiely (Strader et al, 1989a). S~.v.dics by Strader el al. (1989a)

206 further sugge:,! t, ; t the serines in TM5 are important only for ,~goni z ~,nding and their mutation to alaninc does not alter afJtagonist binding. Site-directed mutagenesis of corresponding residues in the a2-adref~ceptor (Wang et al., 1991) suggests that there may be differences in the way catecholamines bi~:3 in ti~e a~- and /32-receptors. Unlike the !32-receptor, serine mutations in TM5 of the t~2-receptor can still exhibit a :'naximal cAMP inhibition with higher agonist concentrations despite a reduced receptor affinity. In the case of the /32-receptor, simgar mutations result in a complete loss of cAMP effects. Further, the two series (204 and 207) in TM5 of the /32-receptor appear equally necessary for agonist affinity, while in the az-receptor the serines (200 and 204) are asymmetrically important, with serine 204 being primary in forming hydrogen bonds with the parahydroxyl group of the catecholamines. In both the a 2and /32-receptors, aspartate 113 of TM3 appears critical for agonist and antagonist binding. Amino acid analysis of the adrenoceptor (a l, a 2 , / 3 , /32) and the dopamine receptors (D~-D 5) demonstrate that the aspartates in TM2 and TM3 and the serines in TM5 are conserved across ~:hese cateeholamine receptors. consistent with a common catecholamine binding pocket (table 1). Yet despite this common structure, other amino acids likely extend into the binding pocket to confer receptor affinity and selectivity. Identifying these residues is particularly important for better understanding of the catecholamine binding pocket and developing more efficacious and selective therapeutic drugs. In addition, as the majority of the information concerning catecholamine receptor structure has been derived from the a 2- and /32-adrenoceptor, it is important to extend these studies to other receptors to determine whether subst~:ntial differences may exist. In the case of the dopamine receptors, only one site-directed mutagenesis study is available in which the effects of aspartate 80 mutations on D 2 receptor binding were evaluated (Ncve et al., 1991). No intormation concerning the aspartate in TM3 or the serines in TM5 of the dopamine receptors is available. Based on the amino acid sequence of the dopamine D z receptor and computer modeling, two conflicting three-dimensional models of the D 2 receptor have been proposed. One model (Dahl et al., 1991) suggests that dopamine interacts with aspartate 80 of TM2 and asparagine 390 of TM6 with the transmembranes being arranged in clockwise manner when viewed from the extracellular space. A second model (Hibert et al., 19911 suggests that dopamine may interact with asparrate 114 of TM3 and the serines of TM5 (194 and 1971 with the transmembranes arranged in a counterclockwise manner. To further evaluate these models and their relevance to other catecholamine receptors, the present study examines the importance of the asparate

114 by site-directed mutagenesis, altering it to either asparagine or glycine, two neutral amino acids. The roles of serines 194 and 197 in TM5 of the D 2 receptor in agonist and antagonist binding were also examined to determine whether dopaminergic comoounds interact with these amino acid residues. Pharmacological data with dopaminergic ligands (Seeman, 1981) indittldt cate "' . . . . tti~ ~'~ meta-hydroxyl group ot• dopaminergic agonists is primarily important in stabilizing agonist binding, suggesting that the scrine residues (194 and 197) of the D 2 receptor may not be equally important for binding affinity as appears to be the case for the /32-receptor. D 2 receptors with single o~ dual serine (194 and 197) mutations were, therefore, evaluated in the present study. As the chemical structures of dopamine and norepinephrine are nearly identical, differing only in a beta-hydroxyl group present in norepinephrine, we hypothesized that there may be specific amino acids near the critical aspartate (114) in TM3 that might aid in conferring receptor selectivity To test this possibility, two adjacent methionines in TM3 of the D 2 receptor in positions 116 and 117 (table 1) were mutated individually or in combination to leucine (116) and cysteine (1171 to more closely approximate the amino _acid ~equence of the/32-adrenoceptor. Methionine 116 is conserved across all the dopamine receptors ( D r D s ) , but is replaced by leucine in the adrenoceptors (a~, a 2, H~, /32). Methionine in position 117 is conserved in D 2 and D 3, but is replaced by either cysteine (D I and D 5) or leucine (D 4) in other dopamine receptors. It was hypothesized that one or both these methionines may provide steric interference with the side chain OH group of epinephrine and norepinephrine, but would allow dopamine, without this moiety, to bind to the D , receptor.

2. Materials and methods

Z 1. Mutagenesis and expression

Mutant receptors of the human D, were prepared with the oligonucleotide-directed mutagenesis system provided by Amersham. Oligonucleotides (2/t-45 bases) were synthesized, purified on polyacrylamide gels (20%), and annealed to a MI3 single-stranded bacteriophage that contained the entire protein coding region of the human D 2 receptor (1-1627 bp) including the 87 bp addition in the tilird cyto,,;olic loop (D2/3). Briefly, the method involves extending the oligonucleotide with Klenow polymerase in the presence of T 4 DNA ligase to generate a mutant heteroduplex. The non-mutated strand is then selectively removed with exonuclea'~e digestion and filtration, leaving the mutant

2(~7 TABLE 1 A m i n o acid alignments of t r a n s m e m b r a n e three and five o1" the adrenoceptors and dopamine receptors. Arrows indicate the aspartate, methionine and serine residues that have been mutated for the D , receptor. Bnti~ human (H) and rat (R) sequences are l-rovided. ~IX

I

V

I

I

I

A1

(H)

. .CDIWAAVDVLCCTASXLSLC.AISLDRY

...........

EPF~ALFSSLGSFYIPLAVZLVI4YCR

A2A

(H)

. .CEZYLALDVLFCTSSI'VHLCAXSLDRY

...........

0KWYVISSCXGSTFAPCLIMXLY'/VR

A2B B1 B2

(H) (H) (H)

. .CGVYLALDVLFCTSSI~/HLCAXSLDRY . .CEZNTSVDVLGVTASIETLCVIALDRY . .CEFW~SIDVLCVTASXETLCVZ~VDRY

........... ........... ...........

ETWYILSSCXGSFFAPCL~/4GLVIq%R NRAYAIASSWSFYVPLVIMAFVYLR NQAYAIASSI~'SFYV~LVZMVFVYS~

B2

(R)

. .CEFWTSZDVLCVTASVETLCVIAVDRY

...........

NQAYAIASSXVSFYVPL~SR

D1 Di 02 02

(H) (R) (H) (R)

D3

CNIWVAFD~','~CSTAS'rL.-'qLCVISVDRY . . . . . . . . . . . . .CNZ',dVAFDZMCSTAS'rLNLCVXSVDRY. . . . . . . . . . . . .CDZFVTLDVMMCTASTLNLCAISIDRY. . . . . . . . . . . .CDXFVTLD~TASZ~SIDRY ...........

SRTY&XSSSVISF'YXPVAIMIVT'ZTR SR'I"ZAXSSSL'rSFYXPVA.T.MIV'I~ NPAFVVYSSIVSFT',,'PFZVTLLVYIX NPAFVVYSSI'VSFYVPFIVTLLVYZK

(R}

. ,CDVF~TLDV~V/CTASILNLCAISZDRY

D4

(H)

. .CD~VMLCTASXFNLCAISVDRF

........... ...........

NPDFVTYSSVVSFY'~PFGVTVLVYAR DRDYVVYSSVCSFFLPCPL~rr.T~T.YWA

D5

(H)

. .CDVWVAFDI/4CSTASILNLCVISVDRY

...........

NRTYAXSSSLXSFYIPVAIMIVTYTR

.

.

.

strand to regenerate the replicative DNA form that is then subcloned in a pCMV expression vector for transfection into eukaryotic cells. A total of seveJ-~ single amino acid mutations were made: Asp (114) to Asn or Gly, Met (116) to Leu, Met (117) to Cys or Gly, Set (194) to Ala, Ser (197) to Ala. Two dual mutations were also designed: Met (116) and Met (117) to Leu (116) and Cys (117) and Ser (194) and Ser (197) to Ala (194) and Aia (197). All pCMV mutant constructs were sequenced to verify that the mutations were correct.

2.2. Transfection COS-1 cells were grown in Dulbeco's modified Eagle's medium with 10% fetal calf serum and subeultured into 90 mm tissue culture plates (1-1.5 x I06 cells) 24 h prior to transient transfection using a calcium phosphate precipitation procedure (Chert and Okayama, 1987). Each 90 mm plate of cells was transfected with 20/xg of pCMV-D 2 wild type or pCMV-D= mutant DNAs. Plasmid DNAs were added to 0.5 ml 0.25 M CaCI 2 to which 0.5 ml of 2 x BBS (50 mM N,N-bis(2-hydroxy-ethyl)-2-aminoethanesulfonic acid. 280 mM NaCI, 1.5 mM NaHPO a, pH = 6.95) was added. This mixture was allowed to remain at 22°C for 10-20 rain, then slowly dripped onto one 90 mm plate of cells. The cells were then grown overnight at 37°C and 3% CO2, washed twice in Versine and once in medium and allowed to grow for an additional 24 h (37°C, 5% CO 2) prior to harvesting. To assess the efficiency of transfection across experiments, all studies involved the cotransfection of/3-galactosidase that was subcloned into pCMV (10 /zg of plasmid)o Based on /3-galactosidase

assays, transfcction efficiencies using this procedure in COS cells ranged from 10 to 40%.

2.3. Radioligand bblding assays At time of cell harvest, the culture medium was removed and each plate of cells was then washed with 10 ml of ligand incubation buffer. For [3H]raclopride (New England Nuclear, 62.7-83.4 Ci/mmol) this buffer consisted of 50 mM Tris (pH = 7.4. 25°C), 120 mM NaC1, 5 mM KCI, 1 mM MgCI 2 and 0.1% ascorbic acid. This buffer was then replaced by fresh ligand incubation buffer (7 ml) in which the transfected cells were harvested by scraping. For cells harvested for [3H]N-0437 (Amersham, 54.(i-98.0 Ci/mmol) binding studies, the same incubation buffer less the 120 mM NaCI and 0.1% ascorbic acid was used. Prior to homogenization (kinematic polytron), an aliquot ~f the celt suspension was saved for ceh counting. The remaining sample was homogenized, with 200 /~1 of the homogenate added to each incubation tube bringing the total membranc homogenate and ligand volume to 250/.tl per tube. Nonspccific binding for [SH]raclopride and [3H]N-0437 was defined by 1 g M (+)-butaclamol. After a 90 min incubation (22°C), membranes were filtered lander vacuum through glass filters (Schleicher and Scbuelt, No. 32) using a Brandel cell harvester. The filters were then rinsed twice with 3.5 ml of washing buffer (50 mM Tris (pH = 7.5), 120 mM NaCI, 5 mM KCI, I mM MgCI z, 4°C) and counted by liquid scintillation spectrometry. The wash buffer for ['~H]N0437 was the same as above except that it did not contain 120 mM NaCI and 0.1% ascorbic acid. For saturation studies a minimum of eight concentrations

208 of radioligand (10-11.1175 nM) with duplicates were used to determine Scatchard pk)ts and all studies were replicated at least once. Similarly, competition studies used a minimum of eight concentrations of unhtbelled compounds, performed in duplicate and repeated at least once. The following drugs were used to evaluate the affinity of dopamieergic agonists and antagonists: apomorphine, bromocriptine, ( + ) - b u t a c l a m o l , chlorpromazine, clozapinc, dopaminc, LY171555, N P A (Npropylnorapomo:phine), and spiperone. A p o m o r p h i n e , bromocriptine, and dop:~mine were purchased from Sigma (St. Louis, MO, USA), and N P A was purchased from Research Biochemical (Natick, MA, USA). All other compounds were a kind gift from James Woods, University of Michigan. A!I binding data was analyzed with the L1GAND program developed by Munson and Rodbard ( 19801. To assess whether" the mutations altered the selectivity of the dopamine rcceptor, [~2slJCYP (cyanopindoIol, New England Nuclear, 21)t)(! C i / m m o l ) binding was performed on transfccted ceils. [~2s I!CYP binding studies were performed as above cxcept that the incubation and wash buffers consisted of 50 mM Tris (pH = 7.47;6), 12.5 mM MgCI 2, and !.5 mM E D T A . In addition, competition studies were performed with epinephrine (Sigma, St. Louis, MO, LISA) and [3H]N-0437.

2.4. ~-Galatosidase (~-gal) assays Aliquots (500 /xl) of the transfected cell homogenates described above were centrifuged at 15,600 x g for 1 rain and stored on ice. 2 0 0 / z l of the supernatant was then added to an equai volume of /3-gal assay solution (1.34 m g / m l O-nitrophenyl-,8-Dgalactoside, 166 mM 2-mercapteothanol, 51)0 m M N a 2 H P O ~, 1 M KCI, 500 mM MgCI 2) arm transferred to a cuvette. The reaction was stopped by ad0~ng 500 /.tl of 1 M Na2CO.~. Spectrophotometric absorbancc of this yellow reaction product was measurcd at 4t0 nm. Absorbance values of the homogenized sample.q were compared to a standard curve consiructed with commercially available /3-galactosidase (Sigma, St. Louis, MO, U S A ) and units of activity were determined. To calculate the percentage of cells cxr:ressmg /3-gal, 90 mm cell culture plates were rinsed twice in PBS, fixed in a formaldehyde ( 2 . 2 % ) / g l u t a r a l d e h y d e (I).2c;~) solu-

1 2 3 4 5 6 7 8 9 28S--

18S-Fig. 1. Northern analysis comparing COS-| cells transfccted with the D, receptor mutants. The lanes are as follows"(1) Aspll4-Asn, (2) Aspll4-Gly, (3) Metll6-Leu, (4) Metll7-C~.';. (5) Melll7-Gly, (6) Serl04-Ala, (7) Ser197-Ala, (8) Scr194 and Ser197-Ab!. (9) Met116Leu and Met117-Cys.

tion for 5 min at 22°C and treated with /3-gal (1 m g / m l ) tbr 1-12 h at 37°C. Cells positive for/3-galactosidase yielded a blue color.

2.5. Northern analysis To aid in determining whether the receptor mutants were expressed, m R N A was extracted from cells transfected with thc D , wild type and mutant D 2 receptors and Northern analysis was performed. R N A samples from transfected COS-1 cells were extracted with 4 M guanidium isothiocyanate and resuspend~d in 5()% formamide, 20 m M m o r p h o l i n e p r o p a n e s u l f o n i c acid (MOPS, pH 7.0), 5 m M sodium acetate, 1 mM E D T A , and 2.2 M formaldehyde. T h e R N A was dcnaturcd at 65°C for ll) rain and clectrophoresed on a 1% agarose gel containing 2.2 M formaldehyde, 2(1 m M M O P S (pH 7.It), 5 m M sodium acetate, and 1 m M E D T A . The R N A samples wcre passively transfcrrcd to Nytran m c m b r a n c s (Schlcicher and Schuell) with 10 x SSC (300 mM NaCI, 30 m M sodium citrate, pH 7.2) and baked for 2 h at 8¢)°C. Thc m e m b r a n e s were then prchybridizcd in 50% formamide, 5 x SSC, 5 x Dcnhardt's, 50 m M sodium phosphate (pH 6.5), and 11.5% SDS for a mininmm of 2 h at 42°C. The hybridization buffcr was thc same as thc prehybridization buffer except that it contained 1 x Denhardt's, 20 m M sodium phosphate, and 10% dextran sulfate. A random-primcd 32p-labelled fragmcnt ( B s t E i l / K p n l ) o f thc human D 2 receptor was used in the hybridization

TABLE 2 Kj values (nM + S.D.) of [~H]raclopride. Asp114-Asn

Asp114-G!y

Mei116-Lcu

Met117-Cys

Mett 17-Gly

Set 194-Ata

Ser197-Ala

Wild type

NSB

NSB

(). 1811 ) + 0.07~

11.124 ( + 0.01)5)

0.211 ( + 0.024)

0.235 ( ± 0.020)

11.714 ,L ( + 0.044)

0.166 ( _+0.015)

Aspartate mutations in position 114 failed to show specific binding (NSB). Analysis of variance of the remaining mutants indicated that there was a difference between mutants anti wild type D 2 ( F = 106515. P < 0.0001). Post-hoe Scheffe comparisons suggested that the serine It17 mutation significantly reduced C' P < 0.05) the affinity for raclopride.

209 TABLE 3 K d Values (riM ± S.D.) of [ 3tt 1N-0437. Asp114-Ash

Asp114-Gly

Met116-Leu

Moll 17-Cys

Metl 17-Gly

Ser194-Ata

Ser197-Ala Ser194-Ala Sert97-Ala

Wild type

NSB

NSB

0.771 ( ± 0.244)

(I.906 ( ± 0Jig0)

1.14 ( ± 11.39)

~).926 ( ± (1.2641

0.906 ( +_0.118)

1),880 ( ± 0.096)

NSB

Aspartate 114 mutations and dual serinc 194 and 197 mutations demonstrated no specific binding (NSB). Other mulations failed to produce any differences in binding compared to the wild type D z,

buffer to probe the filter overnight at 42°C. The blot was washcd once in 2 x SSC and 0.5% SDS at 22°C for ~'5 min, transferred to 0.1 x SSC and 0.5% SDS at 65°C for 30 min and apposed to X-ray film at 22°C for t0 rain. In all of the above procedures, thc D~ mutants and the wild type human D~ receptor were assessed in parallel. The data wcrc analyzed by either one- or two-way A N O V A s and Schiffc post-hoe comparisons were performed.

3. Results

3.1. Ligand binding The affinity of [3H]raclopride for the wild type D 2 varied (0.17-0.55 nM) with tritiated ligand shipments at~,d in all the following experhnents the wild type D~ was evaluated with the D , mutants using the same shipment of tritiatcd ligand. The agonist [3H]N-0437 also failed to consistently demonstrate two binding sites with transiently transfected COS cells. In most cases, the L I G A N D program fit the data best to one site and these are the affinities prescnted in table 3.

3.2. Transfection controls All mutant receptors, when transiently transfected into COS cells, expressed high levels of D~ receptor m R N A (fig. 1). While m R N A levels are not necessary reflections of protein levels, the qualitative differences observed between mutations are not likely due to a lack of receptor transcription. T h e m R N A levels in lanes 6 and 7 are somewhat lower c o m p a r e d to the o t h e r lanes of the Northern blot; however, these results are not consistently observed across experiments. Similarly, co-transfection results with /3-gal that is measured in eve~' experiment suggest that transfection efficiency was equivalent across mutant and wild type D e receptors (data not shown).

3.3. Aspartate (114) mutations Mutation of the negatively charged aspartate (114) residue in TM3 to either asparagine or glycine pro-

duced a dramatic loss of binding affinity for both agonists and antagonists. Saturation studies with both [~H]raclopride (tab!e 2) and [3H]N-0437 (tz~ble 3) failed to demonstrate a n j consistent dopaminergic receptor binding with a s p a n a t e nmtations to either asparagine or glycine, These effects have been replicated over a series of studies and suggest that the negative charge of this aspartate is critical for D e receptor binding.

3.4. Serine 194 attd 197 mutations Individual mutation of serines 194 and 197 in TM5 to alan±he produced asymmetrical effects on dopamine receptor binding. Saturation studies with antagonist [3H]raclopride (tabie 2) and competition studies with [3H]N-0437 (table 4) suggest that serine 197 may be differentially important for dopaminergic binding. As can be seen from table 2, elimination of the potential hydrogen bonds with the serine 197 mutation produced a 4-fold reduction in [3H] ~aclopride binding affinity. In contrast, similar mutations of serine 194 to alanine had no effect on raclol=ride affinity. Surprisingly, individual serine mutations in position 194 and 197 had no effect on the binding affinity of agonist [3H]N-0437 as determined by saturations studits (table 3). Interestingly, however, competition studies using this ligand suggcst that this is not the case for other dopaminergic agonists (table 4). While the loss of dopamincrgic binding affinity varied with agonist, all

TABt.E 4 K, values (nM ± S.D.) of dopamine receptor agonMs competing with [ 3H]N-0437. Serlq4-Ab

Scr197-Ala

Wild type

Apomorphine 5.01 _+ 1.48 89.0 ± 17.0 '~ 2.78± 0.43 Br(mlocripline 1.6(t± 0.19 3.45 ± (I.6 " 1.82_+ (1,31 Dopamine 488.0 ±134.0 1,270,0 ±270.0 ~' 152.0:5:36.t] LYt71555 114.0 ± 52.¢1 1.231k0 ±44(1.0" 217.0 ±14.0 NPA 2.18± 0.57 39,6 ± 3.6" 0.82_+ 0.!0 Two-way ANOVA indicated there were .-dgnificant differences for drug (F = 60.2, P < 0.00011. serinc mutatkm (F = 63.1, P < 0.0001) and drug',